In this study we use the multiple indicator dilution technique to outline the kinetic mechanisms underlying the uptake of rubidium, a cation which, in the steady state, is concentrated by hepatic parenchymal cells. We inject a mixture of (51)Cr-labeled red blood cells (a vascular reference substance), (22)Na (which is confined to the extracellular space, the expected extravascular distribution space for rubidium, in the absence of cellular uptake), and (86)Rb into the portal vein and obtain normalized outflow patterns, expressed as outflowing fractions of each injected mass per milliliter vs. time. The labeled red cell curve rises to the highest and earliest peak and decays rapidly. That for labeled sodium rises to a later and lower peak, and decays less rapidly. Its extrapolated recovery is equal to that for the red cells. The observed (86)Rb curve consists of two parts: an early clearly defined peak of reduced area, related to the (22)Na peak in timing; and a later tailing, obscured by recirculation, so that total outflow recovery cannot be defined (even though it would be expected to be the same). We model the concentrative uptake of (86)Rb and find two corresponding outflow fractions: throughput material, which sweeps past the cell surface as a wave delayed with respect to the vascular reference (tracer which has not entered cells); and exchanging material (tracer which has entered cells and later returns to the circulation). We find that the outflow form of the rubidium curve, the presence of both a relatively clearly defined throughput component and a relatively prolonged low-in-magnitude tailing, is consequent to the concentrative character of the transport mechanism, to the presence of an influx rate constant many times the efflux rate constant. The modeling which we develop is general, and has potential application in situations where transport is nonconcentrative.